United States Environmental Protection Agency Hazardous Waste Engineering Research Laboratory Cincinnati, OH 45268 Research and Development EPA/600/S2-85/031 May 1985 Project Summary '/ I ' Co-Firing of Solid Wastes and Coal at Ames: Pulverized Coal A. W. Joensen, J. L. Hall, J. C. Even, D. Van Meter, P. Gheresus, G. Severns, S. K. Adams, and R. W. White The objectives of this research were to conduct an in-depth evaluation of the environmental, economic, and technical aspects of the resource and energy recovery system in Ames, Iowa. The system includes recovery of ferrous metals, preparation, storage, and cofiring of the refuse-derived fuel (RDF) with coal in the power plant owned by the City of Ames to pro- duce electric power. The evaluation period was three years, and this report covers the third year of research. It includes evalua- tions of the refuse processing plant operation, economics of the total system and individual subsystems, flow stream characterization, and per- formance and environmental emis- sions of the suspension-fired steam generator. Data acquired during the first year's evaluation were previously reported in "Evaluation of the Ames Solid Waste Recovery System. Part I- Summary of Environmental Emissions: Equipment, Facilities, and Economic Evaluations" (EPA-600/2-77-205). This Project Summary was devel- oped by EPA's Hazardous Waste Engineering Research Laboratory, Cin- cinnati, OH, to announce key findings of the research project that is fully documented in a separate report of the same title (see Project Report ordering information at back). Introduction The Ames solid waste recovery system is a continuously operating system that processes municipal solid waste (MSW) to produce a shredded RDF that is burned with Iowa-Western coal mixtures in the tangentially fired steam generator in the Ames municipal power plant. This system consists of a nominal 136-Mg/day (150-ton/day) process plant, a 454-Mg (500 ton) Atlas storage bin, pneumatic transport systems, and the power plant boiler. The process plant incorporates two stages of shredding, ferrous metal recov- ery, and an air classification (density) separator. The full report presents results and con- clusions of the third-year effort of an evaluation of the Ames solid waste recov- ery system, including the process plant studies and boiler environmental and ther- mal performance characterizations. The detailed study objectives are listed in the following section. This evaluation is a major research pro- gram funded by the Environmental Pro- tection Agency (EPA) and the Department of Energy (DOE). Project tasks are being performed jointly by the City of Ames, Iowa, the Engineering Research Institute (ERI) of Iowa State University, the Ames Laboratory/DOE, and MRI. The EPA funding was used to provide for all man- power, major field equipment purchases, power plant and process plant modifica- tions, laboratory analyses of process plant stream characterization, and other sup- plies used in the evaluation of both the power plant and process plant. The DOE funding was used to provide laboratory analysis of all field samples procured from the power plant testing. Additional finan- cial support was provided by ERI, the City of Ames, and the American Public Power Association. System Description The Ames solid waste recovery system consists of three major subsystems: the process plant, the Atlas storage bin, and the existing steam generators of the mu- nicipal power plant which were modified ------- to burn RDF. A general flow diagram is shown in Figure 1. The MSW enters the 45.4 Mg/h process line where primary shredding, ferrous removal, and second stage shredding occurs. The RDF pro- duced from the air density separator (ADS) is transported 152.4 m to the 454-Mg Atlas storage bin through a 36 cm diameter pneumatic transport line. Rejects are subjected to further ferrous removal and are then trucked to the municipal landfill. The RDF is reclaimed from the storage bin by four bucket sweeps which drop the material into two infeed conveyors for the pneumatic transport 61 m to the power plant through two, 20 cm diameter pipes. The RDF is injected into the two opposite corner burners of the 35-MW tangentially fired steam generator. The RDF is burned as a supplemental fuel along with the Iowa-Western coal mixture in suspension, and RDF dropout material is burned on a bottom hopper dump grate installed in 1978. The tangentially fired boiler (No. 7) is a Combustion Engineering Company, Type VU-40S steam generator using balanced draft operation with a Ljungstrom regen- erative air heater and an ESP but no economizer. The two-drum unit operates at 5,860.8 kPa and 485°C steam quality and 163,296 kg/h of steam flow. Combustion air is drawn from the upper part of the building, passed through the forced draft fan through the air heater, and enters the furnace through the corner burner assemblies via two main wind- boxes. Flue gases produced by the fuel combustion in the furnace pass over the primary and secondary superheater tube banks through the convection bank, the air heater, and then through the American-Standard ESP and the induced draft fan (both located outside the building). The flue gases are discharged out the 61-m chimney or stack. Boiler pump discharge feed water is used for superheated steam temperature control, and this spray water is injected between the primary and secondary superheater sections. Process Plant Paniculate Emissions and Dust Evaluation Particulate emissions from the roof ven- tilators of the refuse plant were evaluated by EPA Method 5 particulate sampling techniques. Extensions were added to each of the roof ventilators on the refuse plant to facilitate the samplng. Twenty- four sampling locations on each of two perpendicular traverses across the diameter of the roof ventilator ducts were used, for a total of 48 sample points. At each sample point, the sampling train was operated for 3 min, meaning that a total sample was collected over 144 min of operation. The amount of particulate col- lected was then determined on both a volume and a time basis. The results are reported later in this summary under Power Plant Emission Characterization. In addition to sampling emissions from the roof ventilators, high volume ambient air samplers were placed in the plant to determine the dust concentration at specific locations. The ambient air in the refuse processing plant was sampled at three general eleva- tions in the plant by means of high vol- ume samplers modified to contain 10 cm (4 in.) diameter quartz fiber filters on which the particulate matter collected as the sample train operated. Each sample train was operated for 15 min. The weight of sample was then determined for the time span of the test and recorded for each location in the plant. The three levels sampled in the plant in- cluded the floor level in the general vicini- ty of the first and second stage shredders Municipal Solid Waste Shredder Ferrous -*- Non-Ferrous Separation Rejects <- Non-Ferrous -<- Aluminum -^ Heavy Rejects Air Density Separation (Air Classification) Air Combustible Refuse Derived Fuel Ames Municipal Power Plant Stoker Fired Boiler - 7.5 MW Stoker Fired Boiler - 12.5MW Pulverized Coal Fired Boiler - 35 MW RDF Iowa and Colorado Coal Figure 1. Flow diagram of the Ames solid waste recovery system. 2 ------- and below the air density separator (ADS). The mid-level location sampling was adjacent to the ADS and underneath the bucket elevator in the processing plant. Upper level samples were taken at a walkway in the plant and at the top of the bucket elevator. Following the weighing of the samples, several representative filters were analyzed to ascertain the typical elements present in the dust and the amounts. The amount of dust in the ambient air of the process- ing plant and the results of the trace ele- ment analysis are also presented under Power Plant Emission Characterization. Power Plant In this study, it was determined that two major factors could be controlled at various levels. These factors were the steam generator load, based either on steam flow generated or megawatts of power generated, and the amount of RDF, based on heat energy input to the boiler. The levels chosen were 60, 80, or 100% nominal steam generator load, and 0, 10, or 20% RDF. To obtain sufficient data for statistical analysis, a factorial ex- perimental design with three replications was devised for the steam generators, as summarized in Table 1. The statistical design was a 3 by 3 (three loads, three values of EOF, and three replications) full factorial experiment with 27 runs. To as- certain compliance with Iowa's Envi- ronmental Quality rules, additional mis- cellaneous testing was done. During these tests, the location of the RDF injection point was changed. To satisfy the objectives of the en- vironmental emission study, all ap- propriate input and output streams associated with the operation of the steam generator unit were sampled. A block diagram showing the sample loca- tions of entering and leaving streams is in- cluded as Figure 2. The tests on unit No. 7 are summarized in Table 1. All inputs to and outputs from the steam generator were evaluated, including fuel, combus- tion air, bottom ash, steam, fly ash, and stack gas. All the sampling was con- ducted on a regular basis except the organic species, which were sampled on intermittent days as manpower, instru- mentation, and equipment would allow. Economic Evaluation For 1976, 1977, and 1978, total annual expenses remained relatively constant in that the decreasing principal and interest were approximately balanced by the in- Table 1. Test Matrix for Unit No. 7 Experimental Runs Percent """»» nur Load 60% 80% 100% 80% (Wyoming coal) 100% (Wyoming coal) 100% (Wyoming coal) 0% 3 runs (1976) 3 runs (1976) 3 runs (1976) 3 runs (1978) 3 runs (1978) 10% 2 test runs (1977) - 3 runs (1978) 3 runs (1978) 20% - - 3 runs (1978) 3 runs (1978) Compliance resfs8 4 runs (1978) "RDF injection nozzles relocated to below the coal injection nozzle. creasing operating and maintenance costs. Table 2 shows the relative percentages of operating and maintenance costs al- located to salaries, contractual expense, commodities, and principal and interest. Contractual expenses were higher than salaries during all three years. Principal and interest accounted for nearly half of the operating and maintenance costs. Total operating cost per megagram aver- aged $27.82 in 1976 and $23.87 in 1977. Net cost per megagram averaged $12.47 in 1978, compared with $15.73 in 1976 and $12.61 in 1977. These data are sum- marized in Table 3. Process Plant Emission Characterization The paniculate effluent from the ven- tilator ducts on the refuse plant and the particulate in the ambient air of the pro- cess plant were sampled by appropriate methods. Table 4 summarizes the particulate emissions from the three roof ventilators in operation during this study. Over the 144-min sampling period, the sampling train filters collected particulate effluent in the amounts shown. It should be noted that the emission levels were very low and, in fact, were virtually invisible to observers. Power Plant Emission Characterization The average of heating values, ultimate analysis, and trace elements analysis for both coal and RDF used during the tests on unit No. 7 are as follows: The ash con- tent of the RDF was higher than that of the coal used during 1978, and both the heating values and the amount of sulfur in the RDF were lower than that of the coal for the comparison runs made during this study. The significance of these observa- tions is that as the amount of refuse used in the boiler unit is increased an increased amount of ash will be generated due to the use of refuse. The additional amount of ash was expected to show up partially as fly ash and partially as bottom ash. Consequently, as the RDF increased, the amount of particulate emissions was ex- pected to increase. This was also in agreement with the previous data ob- tained on traveling grate stoker unit Nos. 5 and 6 during 1976 and 1977 studies on the traveling grate units. Because the sulfur content in the RDF was lower than that in the coal, it was also expected that the oxides and sulfur emitted from the smokestack would de- crease significantly with increases in RDF. Based on the tabulation of trace ele- ments from the fuel samples, the RDF contained significantly more copper, lead, titanium, and zinc than coal used as a fuel. As a consequence, the emissions of these four elements were expected to in- crease significantly. The most important of the three elements would be the lead because of its toxicity. Therefore, some additional ambient air sampling was per- formed on a random basis during the 1978 experiments. Germanium, iron, and sulfur , were found in smaller concentrations in the RDF than in the coal, but there did ------- Flow Rate Ultimate Analysis Heating Value Chemical A nalysis & Trace Elements Ash Softening Temperature Filter Particulate Trace Elements Impinger Water Trace Elements Emission Rates of Particulate Particulate Trace Elements Impinger Water Trace Elements Emission Rates of Particulate and Gaseous Species Particulate Sizing Humidity Barometer Intake Temperature Volume Flow Density Ultimate Analysis Heating Value Chemical A nalysis & Trace Elements Ash Softening Temperature Flow Rate Chemical Analysis & Trace Elements Softening Temperature Flow Rate Chemical Analysis & Trace Elements Softening Temperature Figure 2. Sampling locations. Table 2. Solid Waste Plant Operating Expense Distribution, 1976-1978 Operating and maintenance Principal and Labor Contractual Commodities interest Year Total 1976 1977 1978* Average 13.91 18.91 16.73 29.16 24.40 29.25 53.52 6.50 9.21 9.98 50.43 47.48 44.04 46.48 100 100 100 100 "January to September data only. Tabl«3. Total Operating Expenses and Revenues for the Ames Solid Waste Recovery Processing Plant, 1976-1978 Year 1976 1977 1978" Total operating expense ($) 1,033, 186 1,047,734 784,740 Total revenue 1$) 448,721 494,309 411,190 Total net cost 1$) 584,465 553,425 373,550 Refuse processed IMg) 37, 137 43.890 29.958 Total cost/Mg t$/Mg) 27.82 23.87 26.19 Net cost/Mg l$/Mgl 15.73 12.61 12.77 'January through September only. not appear to be a significant difference between RDF and coal in the relative amounts of the other elements. The uncontrolled emissons generally in- crease with RDF except for the 100% load data using coal only. Otherwise all the runs show significant increases in par- ticulate emissions as the amount of RDF increases. It is also apparent from this plot that the initial data obtained using coal only on this boiler in 1976 and 1977 indicate a reverse trend in terms of par- ticulate emissions. The expected par- ticulate would be higher at 100% load than at 60% load as was the case for the 1978 data. This reversed trend in the 1976 data is believed to be related to difficulties in operation of the paniculate collector on unit No. 7 during 1976 and 1977. How- ever, it should be emphasized that the scale for the emissions is significantly ex- panded and that all the emissions with 100% coal as the only fuel are within 2.8 to 3.9 g/MJ of heat energy input. ------- Table 4. Ames Refuse Processing Plant Paniculate Emissions Roof ventilator Roof ventilator EPA Run No. 350 351 352 353 32.16 29.77 37.19 23.26 mg/std m3 18.94 36.14 53.80 31.35 20.85 25.27 31.14 9.798 Average standard deviation Total 71.95 91.18 122.13 64.41 87.42 ±25.74 0.516 0.422 0.518 0.343 g/s 0.209 0.453 0.629 0.368 0.449 0.538 0.632 0.201 Total 1.174 1.413 1.779 0.912 1.320 ±0.368 The controlled emissions generally in- creased with increases in RDF. This result was expected, since the amount of ash in the RDF was proportionally larger than that in the coal. For the 100% coal runs (0% RDF), the decrease in the emissions at 80 and 100% load for the 1978 data was a result of the repair of the ESP late in 1977. Difficulty was experienced with the ESP during 1976 and 1977. Some of the plate retainers in the ESP had failed, which rendered them ineffective during the test runs in 1976 and 1977. This is one reason why the emissions for the 60, 80, and 100% loads in 1976 and 1977 appear to be significantly higher than the emis- sions for the corresponding loads in 1978. Thus, the data obtained in 1978 are much more representative of the usual perform- ance of unit No. 7. Furthermore, the data of 1978 show very consistent trends in the direction anticipated based on the fuel in- put analysis. The effect of RDF on ESP collector effi- ciency drops consistently with increases in RDF. These trends were very consistent for the data obtained in 1978 and showed the ESP efficiency to be higher at 80% load than at 100% load. The effect of the repair between 1977 and 1978 data is also apparent in this figure. For example, for the coal-only runs, the collector efficiency increased from 93.4 to 94.4% at 100% load and from 94.9 to 96.8% at 80% load, thus demonstrating the effect the repair of the ESP had on its performance. The oxides of sulfur (SOX) emitted from the boiler decreased significantly with in- creases in RDF. This decrease amounted to about 50% for 80 and 100% boiler loads in going from 0 to 20% RDF. Thus, an advantage of using RDF with coal is that relatively high-sulfur coal can be used and EPA standards can still be met. The oxides of nitrogen (NOX) generally decreased with increases in RDF at all boiler loads. The decrease was in the range of 10 to 20% and was somewhat dependent on boiler load as the RDF was increased up to 20%. The NOX emissions generally decreased less for the 1978 data than for the 1976-1977 data. This might represent better operation of the boilers and better control of the combustion zone temperatures for the experimental runs of 1978. Except for the 100% load, 20% RDF data point, the chloride emissions for the suspension-fired boiler increased linearly and significantly with increases in RDF. The boiler experienced as much as a ten- fold increase in chloride emissions as the RDF increased from 0 to 20% for all boiler loads in 1978. The chlorides in the stack emissions are believed to have come from the chlorinated hydrocarbons in the RDF. The chlorides dropped in the 1976-1977 data because of the dropout of RDF into the bottom hopper; the bottom grates had not yet been installed. A series of 19 trace elements were sampled from all input and output streams associated with the operation of steam generator unit No. 7. Table 5 lists the trace elements detected in the input fuels of coal and RDF used during the test. The elements selected for analysis are listed by rank order, and the ranking was deter- mined by the concentration given in parts per million (ppm). The standard deviations are also listed. Another column shows the amount of the trace element listed on the basis of mass per unit of energy input to the boiler. The values listed in Table 5 are overall averages for both coal and RDF. The trace elements with higher propor- tions of concentration in coal than in RDF are identified in this table as strontium, beryllium, nickel, and germanium. The elements that were not detected based on the detection limit of the analytical in- strumentation are also indicated. Elements relatively high in concentration in the RDF were zinc, lead, copper, manganese,and vanadium. Conclusions The major result of this project is that RDF can be burned successfully when combined with coal in both stoker-fired and suspension-fired boilers to produce electric power. The net cost per mega- gram to produce this RDF fuel was $15.73 in 1976, $12.61 in 1977, and $12.47 in 1978. The yearly reduction in these net costs was due to plant improvements and increased value of the energy contained in RDF. Major improvements which were made in the Ames solid waste recovery system Table 5. Trace Element Content of Coal and RDF Used as Fuel in Boiler Unit No. 7 Coal RDF Lever* Levef Element Strontium1' Vanadium Manganese Zinc Berylliumf Lead Tin Chromium Nicker* Copper Germanium1' Gallium Antimony Selenium Thallium Mercury Arsenic Cadmium Cobalt ppm 86±28 83±16 76±23 66±41 37± 12 36±13 20±5 19±7 18±5 15±3 5.3 ±0.9 2.5±0,5 BDL BDL BDL BDL BDL BDL BDL ng/J 2.92 ±1.15 2.92± 1. 15 2.92±0.52 2.39± 1.46 1.55±0.49 1.26±0.48 0.71 ±0.17 0.74±0.35 0.63±0.17 0.52±0.08 0.19±0.04 0.09±0.02 BDL BDL BDL BDL BDL BDL BDL Element Zinc Lead Copper Manganese Vanadium Strontium Chromium Tin Antimony Gallium Nickel Selenium Cadmium Germanium Thallium Mercury Arsenic Beryllium Cobalt ppm 763±345 613±289 572±854 194 ±47 154 ±32 46±11 34±8 27 ±8 25±17 16±3 14±4 8±1 6.4±8.1 1.7±0.3 BDL BDL BDL BDL BDL ng/J 4.65±2.13 3.89 ±2.27 3.5815.74 1.18±0.33 0.94±0.19 0.28±0.04 0.28±0.04 0.17±0.06 0.15±0.11 0.10 ±0.02 0.09±0.03 0.05±0.01 0.04±0.05 0.01 ±0.00 BDL BDL BDL BDL BDL Note: BDL signifies the element is below the analytical instrumentation detection limit. "Values listed are overall averages for the coal and RDF used during 1978 tests. ''Trace elements with higher proportions in coal than in RDF. ------- during the three-year comprehensive study are as follows: Addition of dump grates to the 35-MW suspension-fired steam gen- erator Relocation of RDF fuel input nozzles on the 35-MW suspension-fired steam generator to feed RDF below instead of above the coal nozzles Addition of a grit removal system at the processing plant to improve the quality of the RDF Addition of a dust control system at the processing plant to decrease the occurrence of failure of electric motors and mechanical equipment, as well as to improve the worker en- vironment Addition of two crew conveyors at the Atlas storage bin to allow two pneumatic transport lines to pick up RDF from all four drag conveyors and thus reduce the speed of the pull-ring buckets and the wear on the storage bin floor Replacement in 1979 of mechanical collectors with new ones for emis- sions control of the two stoker-fired steam generators to meet environ- mental regulations and permit cofiring of RDF and coal in the stoker boilers The study of boiler performance showed the necessity to improve the quality of RDF in order to reduce slagging and increase boiler performance. A grit removal system was added in the process- ing plant which achieved a 24.5% in- crease in heating value of the RDF, from 11,408.7 to 14,209.1 kJ/kg, and a 54.5% decrease in ash, from 20.99 to 9.55%. Additionally, the boiler fouling impact of RDF was reduced. The addition of the dump grate was the most significant change. This facilitated the successful cofiring of RDF with coal in the suspension-fired steam generator. The relocation of the RDF injection nozzle to a point below the coal injection was found to be important in its effect on lowering emissions. Suspension-fired boiler efficiency decreased 3.3 percentage points when operating at 80% steam load with 20% heat input from RDF and decreased 1.33 percentage points at 100% steam load with 20% heat input from RDF. This decrease was attributed to an increase in moisture loss when RDF is fired. Some furnace slagging was encountered during the period prior to installation of the grit removal system but was reduced after the quality of the RDF was improved by the addition of the grit removal system. Stack paniculate emissions increased slightly with corresponding increases in RDF as a fraction of the fuel input and was due to the presence of lighter RDF particles and increased mass flow. Stack particulate emissions decreased after the RDF injection nozzle was relocated below the coal burners. Oxides of nitrogen (NOX) and oxides of sulfur (SOX) both decreased while chlorides increased with an increase in RDF burning. No discernible trends within the data scatter were noted concerning formaldehyde or hydrocarbon emissons. Increased emissions of the trace elements zinc, copper, lead, and gallium cor- responded to increases in RDF. The two stoker-fired boiler units, used as backup to burn RDF, were modified with new mechanical collectors in 1979. These units previously had difficulty meeting particulate emission standards while only firing coal. Subsequent tests conducted on these units by the City of Ames indicated that coal plus RDF can be successfully burned and meet particulate emission standards as a result of this modification along with the grit removal system at the process plant. A. W. Joensen, J. L Hall, J. C. Even, D. Van Meter. S. K. Adams, P. Gheresus, G. Severns, andR. W. White are with Iowa State University, Ames, IA 50011. Michael Black is the EPA Project Officer (see below). The complete report, entitled"Co-Firing of Solid Wastes andCoal at Ames: Pulverized," fOrder No. PB 85-183 044/AS; Cost $28.00. subject to change) will be available only from: National Technical Information Service 5285 Port Royal Road Springfield. VA 22161 Telephone: 703-487-4650 The EPA Project Officer can be contacted at: Hazardous Waste Engineering Research Laboratory U.S. Environmental Protection Agency Cincinnati, OH 45268 &U. S. GOVERNMENT PRINTING OFFICE:1985/559 111/10848 ------- United States Environmental Protection Agency Center for Environmental Research Information Cincinnati OH 45268 Official Business Penalty for Private Use $300 OCQC329 PS U S ENVIR PROTECTION JGENCY REGION 5 LIBRARY 230 S DEARSCRN STREET CUCAGO IL £0604 ------- |